Alkali vapor-cells have been used extensively since the 1960s in the study of atom-light interactions. Vapor-cell applications, both proposed and realized, include atomic clocks, communication system switches and buffers, single-photon generators and detectors, gas-phase sensors, nonlinear frequency generators, and precision spectroscopy instrumentation. However, most of these applications have only been created in laboratory settings.
Macroscale vapor cells are widely used in macroscale atomic clocks and as spectroscopy references. Macroscale vapor cells are typically 10-100 cm3 in volume, which is insignificant for m3 scale atomic clocks, but far too large for chip-scale atomic clocks which are at most a few cm3 in volume.
A key driver has thus been to reduce vapor-cell size. Traditional vapor-cell systems are large and, if they have thermal control, have many discrete components and consume a large amount of power. To realize the full potential of vapor-cell technologies, the vapor-cell systems need to be miniaturized.
Chip-scale atomic clocks and navigation systems require miniature vapor cells, typically containing cesium or rubidium, with narrow absorption peaks that are stable over time. Miniature vapor cells, and methods of filling them with alkali metals, have been described in the prior art.
However, miniature vapor cells have higher surface-area-to-volume ratios than macroscale vapor cells. It is difficult to load a precise amount of alkali metal into a miniature vapor cell. Furthermore, the amount of alkali vapor in a vapor cell changes over time as the vapor adsorbs, diffuses, and reacts with the walls. Alkali metal vapor pressure may be changed with a small set of known technologies (see Monroe et al., Phys Rev Lett 1990, 65, 1571; Scherer et al., J Vac Sci & Tech A 2012, 30; and Dugrain, Review of Scientific Instruments, vol. 85, no. 8, p. 083112, August 2014). However, these systems are slow, complex, and/or have a short longevity.
A number of patents and patent applications discuss miniature vapor cells and methods of filling them with alkali metals. See U.S. Pat. No. 8,624,682 for “Vapor cell atomic clock physics package”; U.S. Pat. No. 8,258,884 for “System for charging a vapor cell”; U.S. Pat. No. 5,192,921 for “Miniaturized atomic frequency standard”; WO 1997012298 for “A miniature atomic frequency standard”; and WO 2000043842 for “Atomic frequency standard.”
Traditionally, alkali metals have been introduced into magneto-optical trap (MOT) vacuum systems via difficult-to-control manual preparation steps, such as manually crushing a sealed alkali-containing glass ampule inside a metal tube connected to the vacuum system via a control valve. See Wieman, American Journal of Physics, vol. 63, no. 4, p. 317, 1995. This approach requires external heating to replenish the alkali metal inside the vacuum system as needed (a slow process with little control over the amount of alkali metal delivered). The manual labor is non-ideal for automated systems or chip-scale devices.
An alternative exists in the now-common alkali metal dispensers, which are effectively an oven-controlled source of alkali metal, whereby the desired alkali metal is released by chemical reaction when a current is passed through the device. While this process automates the release of alkali metal into the vacuum system, it has difficulty in fabrication compatibility with chip-scale cold-atom devices. Further, the timescales required for generating (warm up) and sinking (pump down) alkali are typically on the order of seconds to minutes, and can vary greatly depending on the amount of alkali metal built up on the vacuum system walls. Alkali metal dispensers also emit other gases initially, so they are not a clean source of alkali.
A rapidly pulsed and cooled variant of the alkali metal dispenser has been developed to stabilize the residual Rb vapor pressure in 100 millisecond pump down time, but the device requires large-dimension Cu heat sinks and complicated thermal design (Dugrain, Review of Scientific Instruments, vol. 85, no. 8, p. 083112, August 2014) which may not easily translate to miniaturization.
Double MOTs wherein the first MOT is loaded at moderate vacuum and then an atom cloud is transferred to a second high-vacuum MOT have been implemented on the laboratory scale. Again, these systems require complicated dual-vacuum systems and controls as well as a transfer system to move the atom cloud from one MOT to the other, none of which is amenable to chip-scale integration.
Light-induced atomic desorption (LIAD) is a recent technique that solves some of the long pump-down times by only releasing a small amount of alkali using a desorption laser; however, this method requires preparing a special desorption target in the MOT vacuum chamber. The desorption laser can interfere with the trapping lasers of the MOT (see Anderson et al., Physical Review A, vol. 63, no. 2, January 2001). It also has yet to demonstrate suitable time constants below 1 second.
Thermoelectric stages can be used to regulate the overall temperature of the vapor cell, but this requires the addition of the thermoelectric stages, a temperature sensor and controller, and a significant amount of power (watts) to maintain the entire temperature of the cell at the correct temperature for MOT operation. The effectiveness of this approach will also depend on the overall size of the MOT cell and the efficiency of the thermoelectric stages, limiting the time constants at which the MOT can be loaded and the residual pressure stabilized.
Atom chips use metal traces patterned via lithographic techniques to create magnetic fields involved in trapping populations of atoms. See U.S. Pat. No. 7,126,112 for “Cold atom system with atom chip wall”; Fortagh et al., Rev. Mod. Phys. 79, 235 (2007) Reichel et al., Atom Chips, Wiley, 2011; and Treutlein, Coherent manipulation of ultracold atoms on atom chips, Dissertation, Ludwig-Maximilians-University Munich, 2008. Atom chips typically are implemented as one wall of a vapor cell. Thus they suffer from the same issues—such as slow vapor pressure rate of change and loss of alkali vapor to the walls—as conventional vapor cells.
Draper Laboratory has developed a solid-state ionic concept based on Cs conducting glass; see U.S. Pat. No. 8,999,123 and U.S. Patent App. Pub. No. 20110247942. However, the Draper technology suffers from two critical deficiencies. The Cs conducting glass has low ion conductivity. The implications of this are shown in Bernstein et al., “All solid state ion-conducting cesium source for atomic clocks,” Solid State Ionics Volume 198, Issue 1, Pages 47-49 (2011), in which >1000 V applied voltage and elevated temperature (˜170° C.) are required to change the alkali content on time scales of ˜100 seconds.
Draper Laboratory discloses the use of a backside electrode which contains a metal (e.g. Ag) that is ionized and injected into the solid electrolyte, the metal being different from the desired atoms in the vapor phase of the vapor cell (e.g. Cs or Rb). This contaminates the solid electrolyte and limits the usable lifetime for the device. For example, a source of silver ions may be used to balance the lost Cs+ ions from the ion-conducting glass. This will contaminate the ion-conducting glass over time and limit its longevity. It is also a source of additional complexity in the fabrication process.
Gold has been used as a source and sink of alkali atoms (Squires, “High repetition rate Bose-Einstein condensate production in a compact, transportable vacuum system,” Ph.D. Thesis, University of Colorado, 2008). However, alkali interaction with gold is a surface effect due to the low diffusivity of alkali atoms in gold. The dense FCC crystal structure of gold leads to low diffusivities of alkali atoms. Gold films thus need to be made very thin, but that reduces any enhancement in alkali transport.
What is desired are vapor cells that work with lower voltages (such as 1-100 V), lower temperatures (such as 20-150° C.), and faster time response (such as 1-100 milliseconds). Response times less than 1 second are crucial because cold-atom lifetime is typically less than 1 second. The excess atoms must therefore be removed from the vapor chamber on time scales less than 1 second in order to have any effect on the cold-atom lifetime. A solution is also sought for the initial vapor-cell loading problem as well as the problem of a loss of alkali vapor over time.